Communication Startup Performance Prediction Across Device Configurations
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The challenge of predicting startup performance in communication devices with thousands of configurations is significant due to the impracticality of testing every combination, leading to incomplete test coverage and compromised performance targets.
Innovation Solution
A machine learning-based method constructs a model using a set of configurations and test results to predict startup performance, enabling accurate predictions for untested configurations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If all configurations are tested manually, then test coverage is improved, but testing time and resources increase significantly
Solution Approach 1:
The patent creates a virtual copy of the testing process through machine learning models. Instead of physically testing each configuration, the system trains models on a subset of test data and uses these models to predict outcomes for all configurations, effectively copying the testing results without repeating the physical tests.
Solution Approach 2:
The system performs preliminary testing on a selected subset of configurations to build training datasets. This preliminary action enables the machine learning models to be trained in advance, allowing rapid prediction for remaining configurations without requiring extensive manual testing of each one.
2Productivity
If machine learning model is used for prediction, then testing time is reduced, but model construction complexity increases
Solution Approach 1:
The patent segments the model construction process into distinct phases: data preparation, model training, model validation, and deployment. This segmentation allows each phase to be handled separately with appropriate tools and methods, reducing the perceived complexity while maintaining high productivity.
Solution Approach 2:
The system introduces an intermediary automated framework that manages the complex model construction process. This intermediary layer handles data preprocessing, model selection, training, and evaluation, shielding users from the underlying complexity while delivering efficient testing results.
Data Source
AI summary
Embodiments of the present disclosure relate to prediction of startup performance of communication device. In one aspect, an electronic device may obtain a set of configurations of a communication device and a set of test values of startup performance of the communication device for the set of configurations, and construct a model for startup performance predication by using a configuration in the set of configurations as an input and a corresponding test result in the set of test values as an output. In another aspect, an electronic device may obtain a configuration of a communication device to be tested; and determine a predicted value of startup performance of the communication device by using the configuration as an input of the model. In this way, predication of startup performance for different configurations may be achieved based on limited test results.